Extended range motion vectors

• US 7,499,495 B2
• Filed: 07/16/2004
• Issued: 03/03/2009
• Est. Priority Date: 07/18/2003
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Srinivasan, Sridhar
• Primary Examiner(s): SENFI, BEHROOZ M
• Application Number:
  US10/893,703
• Publication Number:
  US 20050013372A1
• Time in Patent Office: 1,691 Days
• Field of Search: 348/699, 348/413, 348/416, 348/402, 348/700, 348/701, 348/415, 348/409, 348/401, 348/400, 348/390, 348/384, 348/412, 348/410, 348/419, 382/232, 382/236, 382/239, 375/240.16, 375/240.12, 375/240.24, 375/240.27, 375/240.28, 375/240.01, 375/240.26, 375/240.23, 375/240.03, 375/240.17
• US Class Current: 375/240.16
• CPC Class Codes: H04N 19/44 Decoders specially adapted therefor, e.g. video decoders which are asymmetric with respect to the encoder
• CPC Class Codes: H04N 19/513 Processing of motion vectors
• CPC Class Codes: H04N 19/55 Motion estimation with spatial constraints, e.g. at image or region borders

Abstract

Techniques and tools are described for decoding video data having extended range motion vectors. For example, a sequence level syntax element (such as a flag in a header) indicates whether one or more picture level syntax elements are present for one or more frames of the sequence. If so, then a picture level syntax element for a current frame is processed. The picture level syntax element indicates a motion vector range for the current frame, where the motion vector range may be a default range or one of multiple available extended ranges. Additionally, a signed modulus operation by a factor that depends on the motion vector range may be performed when reconstructing the horizontal and vertical motion vector components to assure that the motion vectors are within the motion vector range.

Claims

1. A computer-implemented method of using a video decoder to transform video from a compressed form to a reconstructed form, the method comprising:

receiving, at the video decoder, encoded data in a bit stream for a video sequence; and

with the video decoder, decoding a picture in the video sequence using the encoded data, including;

with the video decoder, decoding a variable length code indicating a motion vector range for a current frame, wherein the indicated motion vector range is selected from among plural extended ranges and a default range each represented with an associated variable length code in a variable length code table; and

with the video decoder, reconstructing one or more motion vectors for blocks and/or macroblocks of the current frame based at least in part on the indicated motion vector range for the current frame, wherein the reconstructing comprises, for a motion vector component (mv), combining a motion vector differential component (dmv) with a motion vector predictor component (predictor) as follows;

mv=(dmv+predictor)smod b;

wherein the indicated motion vector range is from −

b to b−

1, and wherein smod is a signed modulus operation that assures mv is within the indicated motion vector range.

2. The method of claim 1, wherein each of the plural extended ranges has exact horizontal and vertical limits for the range.

3. The method of claim 1, further comprising:

with the video decoder, processing a sequence level syntax element that indicates one or more picture level syntax elements are present for one or more frames of the sequence including the current frame;

with the video decoder, processing a picture level syntax element including the variable length code indicating the motion vector range for the current frame.

4. The method of claim 3, wherein the sequence level syntax element comprises an element in a sequence header for the bitstream.

5. The method of claim 4, wherein the picture level syntax element comprises an element in a frame header for the current frame.

6. The method of claim 1, wherein the indicated motion vector range for the current frame comprises an allowable horizontal range and an allowable vertical range.

7. The method of claim 1, wherein the indicated motion vector range is one of the plural extended ranges, and wherein the reconstructing comprises processing a first motion vector of the one or more motion vectors for the current frame that extends outside the default range.

8. The method of claim 7, wherein processing the first motion vector comprises using the value of the indicated motion vector range to anticipate how many bits represent a differential component of the first motion vector and processing the bits.

9. A computer-implemented method of using a video decoder to transform video from a compressed form to a reconstructed form, the method comprising:

receiving, at the video decoder, encoded data in a bit stream for a video sequence; and

with the video decoder, decoding a picture in the video sequence using the encoded data, including;

with the video decoder, choosing a motion vector range from among multiple available motion vector ranges; and

with the video decoder, reconstructing a motion vector, including computing a signed modulus by a factor that varies depending on the chosen motion vector range, wherein the reconstructing comprises combining a motion vector differential component (dmv) with a motion vector predictor component (predictor) as follows;

mv=(dmv+predictor)smod b;

wherein mv represents a motion vector component for the motion vector, smod represents the signed modulus, and b represents a component motion vector range factor for the chosen motion vector range.

10. The method of claim 9, wherein the multiple available motion vector ranges include a default range and a plurality of extended ranges that are larger than the default range.

11. The method of claim 9, wherein the motion vector differential component is a horizontal differential component (dmv_x), the motion vector predictor component is a horizontal predictor component (predictor_x), the motion vector component is a horizontal motion vector component (mv_x), and the component motion vector range factor indicates a horizontal motion vector range (−

b<sub>—

x to b</sub>_x−

1).

12. The method of claim 9, wherein the choosing comprises processing a sequence level syntax element relating to motion vector range and a frame level syntax element relating to motion vector range.

13. The method of claim 12, wherein the sequence level syntax element comprises an element in a sequence header for a bitstream and the frame level syntax element comprises an element in a picture header for a current frame.

14. A computer-implemented method of using a video decoder to transform video from a compressed form to a reconstructed form, the method comprising:

with the video decoder, parsing a sequence layer syntax element that indicates whether extended range motion vectors are possible for a sequence of frames;

if the sequence layer syntax element indicates that extended range motion vectors are possible for the sequence of frames, then with the video decoder performing additional processing that allows the use of extended range motion vectors for the sequence of frames, wherein the additional processing comprises;

with the video decoder, parsing a picture level syntax element that indicates whether an allowable motion vector range for a current frame is a default range for the sequence of frames or a larger extended range; and

with the video decoder, performing a signed modulus operation to compute a motion vector for a macroblock or block of the current frame as follows;

mv_—

x=(dmv_—

x+predictor_—

x)smod range_—

x

mv_—

y=(dmv_—

y+predictor_—

y)smod range_—

y

if the sequence layer syntax element indicates that extended range motion vectors are not possible for the sequence of frames, then with the video decoder skipping the additional processing that allows the use of extended range motion vectors for the sequence of frames.

15. The method of claim 14, wherein the picture level syntax element is a variable length code indicating the default range or one of plural available extended ranges.